Skip to content

Latest commit

 

History

History
75 lines (51 loc) · 2.23 KB

README_clip.md

File metadata and controls

75 lines (51 loc) · 2.23 KB

Multimodal CLIP Embedding Microservice

The Multimodal CLIP Embedding Microservice provides a powerful solution for converting textual and visual data into high-dimensional vector embeddings. These embeddings capture the semantic essence of the input, enabling robust applications in multi-modal data processing, information retrieval, recommendation systems, and more.

✨ Key Features

  • High Performance: Optimized for rapid and reliable embedding generation for text and images.
  • Scalable: Capable of handling high-concurrency workloads, ensuring consistent performance under heavy loads.
  • Easy Integration: Offers a simple API interface for seamless integration into diverse workflows.
  • Customizable: Supports tailored configurations, including model selection and preprocessing adjustments, to fit specific requirements.

This service empowers users to configure and deploy embedding pipelines tailored to their needs.

🚀 Quick Start

1. Launch the Microservice with Docker

1.1 Build the Docker Image

To build the Docker image, execute the following commands:

cd ../../..
docker build -t opea/embedding:latest \
  --build-arg https_proxy=$https_proxy \
  --build-arg http_proxy=$http_proxy \
  -f comps/embeddings/src/Dockerfile .

1.2 Start the Service with Docker Compose

Use Docker Compose to start the service:

cd comps/embeddings/deployment/docker_compose/
docker compose up clip-embedding-server -d

2. Consume the Embedding Service

2.1 Check Service Health

Verify that the service is running by performing a health check:

curl http://localhost:6000/v1/health_check \
  -X GET \
  -H 'Content-Type: application/json'

2.2 Generate Embeddings

The service supports OpenAI API-compatible requests.

  • Single Text Input:

    curl http://localhost:6000/v1/embeddings \
  -X POST \
  -d '{"input":"Hello, world!"}' \
  -H 'Content-Type: application/json'

  • Multiple Texts with Parameters:

    curl http://localhost:6000/v1/embeddings \
  -X POST \
  -d '{"input":["Hello, world!","How are you?"], "dimensions":100}' \
  -H 'Content-Type: application/json'